Chapter 3. 0 Methodology
3.3 Sampling strategy
As mentioned in the introduction, this research aimed to address some gaps in the empirical research, by gleaning information from a range of professionals within the case study organisation, both in formal leadership, and non-formal (i.e academic roles). The sampling strategy was therefore approached with this in mind, whilst aiming to achieve the overall objectives of the study. As Cohen et al (2003, p92) note, “The quality of a piece of research not only stands or falls by the appropriateness if methodology and instrumentation, but also by the suitability of the sampling strategy that has been adopted”.
In deciding a sampling strategy, researchers may give consideration to two main methods of sampling; probability (also known as random sampling) or non-probability (also known as purposive sampling), (Cohen et al 2003). Appropriate to this research is purposive sampling where a sample can be built to satisfy the researcher’s needs and there is a specific targeting of a particular group, primarily some colleagues in formal leadership roles, as well as academics in a non-formal leadership position within University College. ‘In purposive sampling the researcher applies his/her experience and judgement to select cases which are representative.’ (Briggs et al, 2012, p101).
Whilst the purposive sampling method does not pretend to represent the wider population, it is acknowledged nevertheless that the validity of the research may be compromised by the sample size. It was proposed to identify ten academic staff for interview, and five in formal roles, which would allow for cross comparison and identification of resulting themes. It is acknowledged that a larger sample for interview would add to the validity of the results, but this was not possible within the scope of the study. Approximately two hundred questionnaires were distributed was thought sufficient to produce some useful data, given consideration for realistic response rates (Cohen et al, 2007). The response rate would need to be taken into account when analysing data; caution is to be advised as ‘a lot of defective research results from attempting to extrapolate from tiny samples to grand theory’ (Swetnam, 2000, p43).
When accessing the sample “researchers need to ensure not only that access is permitted, but is, in fact, practicable”. (Cohen et al, 2003, p99). In carrying out research, access may be denied not only because of issues such as sensitivity, but because of very practical reasons such as time constraints. In terms of sensitivity, it is suggested that interviewing prospective leaders is a very different prospect that interviewing established leaders who may have ‘learned to master their emotional vulnerabilities’ (Briggs et al, 2012, p198). It will be important therefore to bear in mind these contextual issues when analysing results from these interviews.
In addition to sensitivity, key ethical issues also needed to be considered including terms of voluntary informed consent and the right to withdraw. In particular “researchers engaged in case study research must consider the extent to which their own reflective research impinges on others, for example in the case of the dual role of teacher and researcher and the impact on colleagues” (BERA 2004, p6).
It was thought appropriate to distribute the questionnaire at the University College Staff conference, held in a dedicated venue a short distance from University College. The event is an annual conference designed to report on key research being undertaken by staff, and on issues affecting the institution, including learning and teaching. As such, all academic staff are offered the opportunity to attend, whilst allowing time for networking.
The time frame of the study coincides with the end of undergraduate teaching, and the completion of official duties of academics such as exam boards, and therefore attracts the majority of academic staff.
In terms of distribution, it was important that the desired sample was reached, whilst making the research practicable. The conference organiser (Head of the Research Department) was approached for consent to include the questionnaire in the conference delegate packs. The research was also highlighted in the opening address of the conference and delegates invited to complete the questionnaire. It is felt that this part aided the response rate of study.
Where response rates are concerned ‘various follow up methods may improve matters but cause delays and could conceivably invalidate results by making a cross sectional study a longitudinal one, that is a study of the same population over time’ (Swetnam, 2000, p60). A college- wide e-mail was sent inviting staff who had not yet done so to complete the questionnaire. However, it was acknowledged that there is the probability that some academics would not complete the questionnaire at all. (Kumar, 2005) suggests there is a self-selecting bias when an ‘open’ invitation is sent to a sample – it may be those that complete the questionnaire have particularly strong attitudes, attributes or motivation. This was considered when interpreting the data.
When considering sampling, overall, ‘what is important is that firstly, we should use the best sample we can within the resources and possibilities available’ (Briggs et al, 2012, p99). In addition, ‘where the best sample falls short of what might have been the ideal, then we should acknowledge this shortcoming,’ (ibid). Whilst it was therefore important to bear in mind that given the reasons outlined, the responses may not be completely representative of the study population, it was anticipated that there would be a sufficient amount of data with which to work. The subsequent process of data collection and analysis will now be explored.